Research On Asynchronous Motor Vector Control System Based On Neural Network

In recent years, with the appearance of the new power electron apparatus and microprocessor and the development of the control theory on the Alternating Current (AC) motor, the technology of variable frequency speed variation has improved rapidly. Vector control is one of the strategies for the AC motor drive with high performance. And it has been applied everywhere in the industries. This paper does some research works of the rotor flux oriented vector control technology.This paper elaborates the development of AC speed regulation system and the research emphasis, then introduces the different control strategies of asynchronous motor speed control system. In this thesis, the basic principles of vector control, vector coordinate transformation are presented. The mathematical model of asynchronous motor and the rotor flux oriented vector control system are discussed, Then the accuracy of flux observation is analysed comprehensively. This paper introduces the current model and voltage model of rotor flux estimation model and puts forward a method based on combination model of rotor flux observer. There are many problems due to the installation of speed sensor, the speed sensorless technology research receives attention. Then the model reference adaptive control systems (MRAS) and the speed estimation methods based on MRAS are discussed.With many advantages, vector control is widely applied for AC motor variable frequency speed control system. But it has many problems, such as complex system structure, nonlinearity and parameter variations. Artificial neural network has strong learning ability, trained multilayer neural network can approach nonlinear function with arbitrary precision. On the basis of analyzing the principle of the neural network, the paper has researched on the applications of neural network in the AC motor drive. Two kinds of neural network speed controller have been studied. One is the single-neuron adaptive PID, the other is BP neural network PID controller, the BP neural network adjusts PID parameters online. Through the study of the two methods, neural network can improve the system anti-interference ability.